# Evaluating (/docs/accelerate/evaluating)

> For the complete Prisma documentation index, see [llms.txt](https://www.prisma.io/docs/llms.txt). A markdown version of any docs page is available by appending `.md` to its URL.

Learn about evaluating Prisma Accelerate

Location: Accelerate > Evaluating

Prisma Accelerate provides connection pooling and global edge caching. Its connection pooler is available in 16 regions and helps applications load-balance and scale database requests based on demand.

Evaluate Accelerate with high volume to see how it performs under load.

## How Accelerate's connection pool optimizes performance under load [#how-accelerates-connection-pool-optimizes-performance-under-load]

Prisma Accelerate employs a dynamic, serverless connection pooling infrastructure. When a request is made, a connection pool is quickly provisioned for the project in the region assigned while configuring Prisma Accelerate. This connection pool remains active, serving many additional requests while reusing established database connections. The connection pool will disconnect after a period of inactivity, so it's important to evaluate Prisma Accelerate with a consistent stream of traffic.

**Key Benefits:**

* **Optimized Query Performance:** The serverless connection pooler adapts to the query load, ensuring the database connections are managed efficiently during peak demand.

  > Prisma Accelerate's connection pooler cannot improve the performance of queries in the database. In scenarios where query performance is an issue, we recommend optimizing the Prisma query, applying indexes, or using Accelerate's edge caching.

* **Maximize Connection Reuse:** Executing a consistent volume of queries helps maintain active instances of Accelerate connection poolers. This increases connection reuse, ensuring faster response times for subsequent queries.

## Evaluating Prisma Accelerate connection pooling performance [#evaluating-prisma-accelerate-connection-pooling-performance]

Below you will find an example of how to evaluate Prisma Accelerate using a sample model:

```prisma
model Notes {
  id        Int       @id @default(autoincrement())
  title     String
  createdAt DateTime  @default(now())
  updatedAt DateTime? @updatedAt
}
```

```typescript
import { PrismaClient } from "@prisma/client";
import { withAccelerate } from "@prisma/extension-accelerate";

const prisma = new PrismaClient().$extends(withAccelerate());

function calculateStatistics(numbers: number[]): {
  average: number;
  p50: number;
  p75: number;
  p99: number;
} {
  if (numbers.length === 0) {
    throw new Error("The input array is empty.");
  }

  // Sort the array in ascending order
  numbers.sort((a, b) => a - b);

  const sum = numbers.reduce((acc, num) => acc + num, 0);
  const count = numbers.length;

  const average = sum / count;
  const p50 = getPercentile(numbers, 50);
  const p75 = getPercentile(numbers, 75);
  const p99 = getPercentile(numbers, 99);

  return { average, p50, p75, p99 };
}

function getPercentile(numbers: number[], percentile: number): number {
  if (percentile <= 0 || percentile >= 100) {
    throw new Error("Percentile must be between 0 and 100.");
  }

  const index = (percentile / 100) * (numbers.length - 1);
  if (Number.isInteger(index)) {
    // If the index is an integer, return the corresponding value
    return numbers[index];
  } else {
    // If the index is not an integer, interpolate between two adjacent values
    const lowerIndex = Math.floor(index);
    const upperIndex = Math.ceil(index);
    const lowerValue = numbers[lowerIndex];
    const upperValue = numbers[upperIndex];
    const interpolationFactor = index - lowerIndex;
    return lowerValue + (upperValue - lowerValue) * interpolationFactor;
  }
}

async function main() {
  const timings = [];

  // fire a query before going to the loop
  await prisma.notes.findMany({
    take: 20,
  });

  // we recommend evaluating Prisma Accelerate with a large loop
  const LOOP_LENGTH = 10000;

  for (let i = 0; i < LOOP_LENGTH; i++) {
    const start = Date.now();
    await prisma.notes.findMany({
      take: 20,
    });

    timings.push(Date.now() - start);
  }

  const statistics = calculateStatistics(timings);
  console.log("Average:", statistics.average);
  console.log("P50:", statistics.p50);
  console.log("P75:", statistics.p75);
  console.log("P99:", statistics.p99);
}

main()
  .then(async () => {
    await prisma.$disconnect();
  })
  .catch((e) => {
    await prisma.$disconnect();
    process.exit(1);
  });
```

## Evaluating Prisma Accelerate caching performance [#evaluating-prisma-accelerate-caching-performance]

Prisma Accelerate's edge cache is also optimized for a high volume of queries. The cache automatically optimizes for repeated queries. As a result, the cache hit rate will increase as the query frequency does. Adding a query result to the cache is also non-blocking, so a short burst of queries might not hit the cache the way a sustained load does.

To evaluate Accelerate's edge caching, modify the script above as follows:

```typescript
import { PrismaClient } from "@prisma/client";
import { withAccelerate } from "@prisma/extension-accelerate";

const prisma = new PrismaClient().$extends(withAccelerate());

function calculateStatistics(numbers: number[]): {
  average: number;
  p50: number;
  p75: number;
  p99: number;
} {
  if (numbers.length === 0) {
    throw new Error("The input array is empty.");
  }

  // Sort the array in ascending order
  numbers.sort((a, b) => a - b);

  const sum = numbers.reduce((acc, num) => acc + num, 0);
  const count = numbers.length;

  const average = sum / count;
  const p50 = getPercentile(numbers, 50);
  const p75 = getPercentile(numbers, 75);
  const p99 = getPercentile(numbers, 99);

  return { average, p50, p75, p99 };
}

function getPercentile(numbers: number[], percentile: number): number {
  if (percentile <= 0 || percentile >= 100) {
    throw new Error("Percentile must be between 0 and 100.");
  }

  const index = (percentile / 100) * (numbers.length - 1);
  if (Number.isInteger(index)) {
    // If the index is an integer, return the corresponding value
    return numbers[index];
  } else {
    // If the index is not an integer, interpolate between two adjacent values
    const lowerIndex = Math.floor(index);
    const upperIndex = Math.ceil(index);
    const lowerValue = numbers[lowerIndex];
    const upperValue = numbers[upperIndex];
    const interpolationFactor = index - lowerIndex;
    return lowerValue + (upperValue - lowerValue) * interpolationFactor;
  }
}

async function main() {
  const timings = [];

  // fire a query before going to the loop
  await prisma.notes.findMany({
    take: 20,
    cacheStrategy: {
      ttl: 30,
    },
  });

  // we recommend evaluating Prisma Accelerate with a large loop
  const LOOP_LENGTH = 10000;

  for (let i = 0; i < LOOP_LENGTH; i++) {
    const start = Date.now();
    await prisma.notes.findMany({
      take: 20,
      cacheStrategy: {
        ttl: 30,
      },
    });

    timings.push(Date.now() - start);
  }

  const statistics = calculateStatistics(timings);
  console.log("Average:", statistics.average);
  console.log("P50:", statistics.p50);
  console.log("P75:", statistics.p75);
  console.log("P99:", statistics.p99);
}

main()
  .then(async () => {
    await prisma.$disconnect();
  })
  .catch((e) => {
    await prisma.$disconnect();
    process.exit(1);
  });
```

## Related pages

- [`Caching queries`](https://www.prisma.io/docs/accelerate/caching): Learn everything you need to know to use Accelerate's global database caching
- [`Compare Accelerate`](https://www.prisma.io/docs/accelerate/compare): Learn how Prisma Accelerate compares to other connection poolers like pgbouncer
- [`Connection Pooling`](https://www.prisma.io/docs/accelerate/connection-pooling): Learn about everything you need to know to use Accelerate's connection pooling
- [`Examples`](https://www.prisma.io/docs/accelerate/examples): Check out ready-to-run examples for Prisma Accelerate
- [`Getting started`](https://www.prisma.io/docs/accelerate/getting-started): Learn how to get up and running with Prisma Accelerate